CN104376389A - Master-slave type micro-grid power load prediction system and master-slave type micro-grid power load prediction method based on load balancing - Google Patents
Master-slave type micro-grid power load prediction system and master-slave type micro-grid power load prediction method based on load balancing Download PDFInfo
- Publication number
- CN104376389A CN104376389A CN201410759321.XA CN201410759321A CN104376389A CN 104376389 A CN104376389 A CN 104376389A CN 201410759321 A CN201410759321 A CN 201410759321A CN 104376389 A CN104376389 A CN 104376389A
- Authority
- CN
- China
- Prior art keywords
- module
- prediction
- data
- load
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention provides a master-slave type micro-grid power load prediction system based on load balancing. The master-slave type micro-grid power load prediction system comprises a main server; the main server is used for carrying out high-complexity mathematical calculation and large-scale data storage; information interaction can be carried out through an Ethernet switch and distributed secondary stations of the prediction system; the distributed secondary stations of the prediction system can transmit tasks which are required to be subjected to large-scale calculation to the main server; the main server performs calculation of the tasks; and each distributed secondary station of the prediction system is used for acquiring real-time data of fans in a micro-grid or a secondary micro-grid, real-time data of photovoltaic generated power and load data of a region. The power and the load of the micro-grid can be predicted precisely, a prediction result provides precise data support for an energy management system and a micro-grid controller, the prediction cost is reduced, and the using efficiency of a server of the prediction system and the using efficiency of a device of the prediction system are improved. Data-level load balancing can be realized on the system, and network element internal threading-level load balancing is realized in each network element.
Description
Technical field
The present invention relates to microgrid power prediction, micro-grid load prediction field, what be specifically related to is a kind of master-slave mode microgrid power load prediction system based on load balancing and method thereof.
Background technology
Micro-capacitance sensor is a kind of novel electric network composition, includes wind-powered electricity generation, photovoltaic, diesel generation, energy storage, load, Control protection module etc.Micro-capacitance sensor can teaching display stand control, the autonomous system of protect and manage, both can be incorporated into the power networks with external electrical network, also can isolated operation.Micro-capacitance sensor fully can promote the efficiency utilization of distributed power source and regenerative resource.
Photovoltaic, wind-powered electricity generation industry develop rapidly in recent years, and increasing low profile photovoltaic generating and wind-powered electricity generation are applied in micro-capacitance sensor field, and accurate wind-powered electricity generation and photovoltaic power prediction contribute to micro-capacitance sensor scheduling controlling and safe operation.If the load data in following certain micro-capacitance sensor can be determined, also there is important meaning to micro-capacitance sensor economic load dispatching and energy management particular moment.
Photovoltaic in the market, wind-powered electricity generation, load prediction system need multiple acquisition channel, and each prognoses system is separate again, therefore causes data redundancy.In addition, the prognoses system of current each family needs to develop the communication of multiple communication interface for prognoses system and outside collector, controller, energy management system, scheduling, once Real-Time Communication Interface interrupts, the input data of prognoses system will produce larger error, also can be inaccurate to predicting the outcome.
Photovoltaic generation and wind-power electricity generation are subject to inside even from weather, exert oneself unstable often.Traditional photovoltaic and wind-powered electricity generation prognoses system need the support of meteorological department's numerical weather forecast, but current numerical weather forecast temporal resolution and spatial resolution generally all cannot reach the requirement of accurately predicting, therefore also increase microgrid energy management system and control, regulate and exert oneself and the difficulty of balancing the load.
Current power prediction and load prediction adopt neural network algorithm mostly, and neural network algorithm training period system generally needs to spend a large amount of resources for calculating neural network weight, and this just needs to use high performance server.In actual motion, the training of neural network is generally that even the longer time just needed training once in one month, and high cost, high performance server do not obtain higher utilization rate.
Summary of the invention
For the deficiency that prior art exists, the present invention seeks to be to provide a kind of master-slave mode microgrid power load prediction system based on load balancing and method thereof, for micro-capacitance sensor provides accurate power and load prediction data, realize system-level and load balancing that is network element internal thread-level, improve the whole utilization efficiency of system and device.
To achieve these goals, the present invention realizes by the following technical solutions:
Based on a master-slave mode microgrid power load prediction system for load balancing, it comprises master server, stores for mathematical computations and data;
The distributed substation of prognoses system, for gathering the relevant real time data of local prediction, the power in prediction local zone and load data, carry out modified weight calculating simultaneously; Described master server too network switch and the distributed substation of prognoses system carries out information interaction,
Described master server comprises with lower module:
Meteorological acquisition module, for gathering the free numerical weather forecast of automatic network;
Accurate weather prognosis module is foundation according to local longitude, dimension, height above sea level, geopotential unit, simultaneously according to the geography information in micro-capacitance sensor garden and building schematic diagram, carries out three-dimensional modeling;
Database, for storing meteorological historical data and image data, and power, the load data in micro-capacitance sensor garden; Remote master server communication module is used for communicating with the distributed substation of each prognoses system, gathers the weather data on Internet simultaneously;
Communication module, based on common ethernet communication mode, uses ICP/IP protocol.
The long-range training module of neural network, for training power prediction and load forecasting model;
Equalization algorithm module, is responsible for asset creation and the Resourse Distribute of control neural network training module;
Remotely predicting module, according to general power and the total load data of the predict the outcome multiple micro-capacitance sensor of prediction or multiple sub-micro-capacitance sensor of the distributed substation of prognoses system, or when breaking down in the distributed substation of certain prognoses system, replaces it to complete power and load prediction.
Further, the distributed substation of described prognoses system comprises:
Central processing unit, for dispatching the collaborative work between each submodule, completes basic system cloud gray model and data processing;
Prognoses system communication module, for the communication before the distributed substation of prognoses system and master server, gathers local environment Monitoring Data simultaneously;
Memory module, for storing the historical data of nearly trimestral environment monitor, load, power;
Neural network algorithm training module, for training power prediction and the online retraining of load forecasting model;
Neural network algorithm training prediction module, is weighted prediction according to the forecast model that neural network algorithm training module trains out;
Equalization algorithm module, be responsible for asset creation and the Resourse Distribute of control neural network training module, when local resource can not meet training requirement or can affect data acquisition, power load prediction, central processing unit can be trained to master server request remote opening, load forecasting model file can be sent to this locality after master server completes training;
Local prediction module, for gathering the relevant real time data of local prediction, the power in prediction local zone and load data;
Described communication module, memory module, neural network algorithm training module are all connected with central processing unit with local prediction module, described equalization algorithm module is all connected with neural network algorithm training module with neural network algorithm training prediction module, described local prediction module, neural network algorithm training module and prognoses system communication module are all connected with memory module, are convenient to the storage of data.
Further, described local prediction module comprises photovoltaic prediction module, wind-powered electricity generation prediction module and load prediction module.
As preferably, the neural network algorithm of the long-range training module of described neural network adopts the BP neural network algorithm improved, and the BP neural network algorithm of described improvement is as follows:
Right value update Δ W
lj:
Wherein, W
ljrepresent the connection weights between l neuron to an output layer jth neuron, O
ijrepresent that neural network exports, y
jrepresent desired value, v
lrepresent l neuron input value, θ
jrepresent threshold value, L represents neuron number.
As preferably, what described database adopted is Oracle or sybase database.
Based on a master-slave mode microgrid power load forecasting method for load balancing, its method step is as follows:
(1) distributed capture prediction steps; Gather the real time data that local prediction is relevant, by the power in Neural Network Prediction local zone and load data, carry out modified weight calculating simultaneously, export budget result;
(2) meteorological acquisition step, for gathering the free numerical weather forecast of automatic network;
(3) weather prognosis step, the locality adopted according to step (1) among a small circle longitude, dimension, height above sea level, geopotential unit is foundation, carries out three-dimensional modeling; The network weather data that model obtains using step (2), as initial conditions, adopt physical equation and the meteorological change of thermodynamical equilibrium equation simulation, and after neural network algorithm correction, finally obtain accurate weather information in micro-capacitance sensor garden;
(4) general power and the total load data of micro-capacitance sensor are predicted; Predicting the outcome and general power through predicting multiple micro-capacitance sensor or multiple sub-micro-capacitance sensor and total load data according to step (1);
Every prediction is all based on neural computing, neural computing is based on BP neural network algorithm, traditional neural network performs neural net model establishing and computing by serialized manner on certain station server or computer, the present invention utilizes " load balance scheduling algorithm ", perform to " remote master server " neural net model establishing of single serial and computing executed in parallel or partial arithmetic job invocation on certain computer or server, about the parallel method of single personal computer or server with reference to figure 5.
The accurate weather prognosis module of remote master server can according to locality among a small circle longitude, dimension, height above sea level, geopotential unit be foundation, simultaneously with reference to the geography information in micro-capacitance sensor garden and building schematic diagram, carry out three-dimensional modeling.Model is using network weather data as initial conditions, use physical equation and the meteorological change of thermodynamical equilibrium equation simulation, after using neural network correction, finally to obtain in micro-capacitance sensor garden accurate weather information within the scope of every 30 × 30 square metres, temporal resolution is less than 5 minutes.
Remote master server database uses the industrial Oracle that generally uses or sybase database.For storing meteorological historical data and image data, and power, the load data in micro-capacitance sensor garden.
Remote master server communication module is used for communicating with the distributed substation of each prognoses system, gathers the weather data on Internet simultaneously.Communication module, based on common ethernet communication mode, uses ICP/IP protocol.
The long-range training module of remote master server neural network is used for training power prediction and load forecasting model.Wherein neural network algorithm adopts the BP neural network algorithm improved.Wherein right value update, the computing formula that we use is as follows:
Wherein W
ljrepresent the connection weights between l neuron to an output layer jth neuron, O
ijrepresent that neural network exports, y
jrepresent desired value, v
lrepresent l neuron input value, θ
jrepresent threshold value, L represents neuron number.
Remote master server remotely predicting module can according to the general power of the predict the outcome multiple micro-capacitance sensor of prediction or multiple sub-micro-capacitance sensor of the distributed substation of prognoses system and total load data, or when breaking down in the distributed substation of certain prognoses system, replace it to complete the function of power and load prediction.
The method of neural network algorithm is: be first initialization weight data, each weights assignment random number, and random number range is between-1 to 1; Using the input value of historical data sample as neural network, calculate the input and output of every one deck neural network respectively.After obtaining the final output of neural network, ask theoretical value and calculated value square error, if error meets pre-conditioned, then train end; If do not meet pre-conditioned, then judge maximum cycle again, if exceed maximum cycle equally also terminate training, otherwise backwards calculation every layer neuron partial gradient is revised neuron weights; Until meet the pre-conditioned of square error or maximum cycle.
First equalization algorithm module can check operating system version, and loads corresponding grand, initialization global variable.Algorithm can be added up how many CPU and start to monitor the occupancy of each CPU.Balance module algorithm can create a task manager and be used for task burst, task monitors, distributing system resource and recovery system resource in internal memory.If when local resource can meet all tasks, task manager can start the task after performing burst.If local resource can not meet all tasks, and this task cannot be waited for, equalization algorithm module can connect master server, is performed by this job invocation to master server.
Each subtask first can initialization task counter, ensure that this task occupying system resources time is limited, will the task of Data import, data processing be completed after this tasks carrying, and externally I/O request reading and writing of files, file lock, by ensureing the uniqueness of file read-write, prevents corrupt data.
The present invention can realize the detailed predicting of microgrid power and load, predicts the outcome as energy management system and micro-capacitance sensor controller provide accurate Data support, reduces forecast cost simultaneously, improves the service efficiency of prognoses system server and device.Can realize the load balancing of data level in system, each network element internal realizes network element internal thread-level load balancing.
Accompanying drawing explanation
The present invention is described in detail below in conjunction with the drawings and specific embodiments;
Fig. 1 is for being system architecture schematic diagram of the present invention;
Fig. 2 is neural network BP training algorithm schematic flow sheet;
Fig. 3 is prognoses system of the present invention distributed substation structured flowchart;
Fig. 4 is present system level load balance scheduling algorithm flow chart;
Fig. 5 is network element internal thread-level load balance scheduling algorithm flow chart of the present invention.
Embodiment
The technological means realized for making the present invention, creation characteristic, reaching object and effect is easy to understand, below in conjunction with embodiment, setting forth the present invention further.
A kind of master-slave mode microgrid power based on load balancing and load forecasting method comprise: the distributed substation of remote master server, prognoses system.
Remote master server includes: meteorological acquisition module, accurately weather prognosis module, database, communication module, the long-range training module of neural network, remotely predicting module, equalization algorithm module etc.
The distributed substation of prognoses system is made up of following a few part: core processor, communication module, memory module, neural network algorithm training module, local prediction module, equalization algorithm module.
The meteorological acquisition module of remote master server is for gathering the free numerical weather forecast of automatic network, and this weather forecast temporal resolution is generally more than one hour, and spatial dimension is approximately the region of more than 6 × 6 square kilometres.
The accurate weather prognosis module of remote master server can according to locality among a small circle longitude, dimension, height above sea level, geopotential unit be foundation, simultaneously with reference to the geography information in micro-capacitance sensor garden and building schematic diagram, carry out three-dimensional modeling.Model is using network weather data as initial conditions, use physical equation and the meteorological change of thermodynamical equilibrium equation simulation, after using neural network correction, finally to obtain in micro-capacitance sensor garden accurate weather information within the scope of every 30 × 30 square metres, temporal resolution is less than 5 minutes.
Remote master server database uses the industrial Oracle that generally uses or sybase database.For storing meteorological historical data and image data, and power, the load data in micro-capacitance sensor garden.
Remote master server communication module is used for communicating with the distributed substation of each prognoses system, gathers the weather data on Internet simultaneously.Communication module, based on common ethernet communication mode, uses ICP/IP protocol.
The long-range training module of remote master server neural network is used for training power prediction and load forecasting model.Wherein neural network algorithm adopts the BP neural network algorithm improved.Wherein right value update, the computing formula that we use is as follows:
Wherein W
ljrepresent the connection weights between l neuron to an output layer jth neuron, O
ijrepresent that neural network exports, y
jrepresent desired value, v
lrepresent l neuron input value, θ
jrepresent threshold value, L represents neuron number.
Remote master server remotely predicting module can according to the general power of the predict the outcome multiple micro-capacitance sensor of prediction or multiple sub-micro-capacitance sensor of the distributed substation of prognoses system and total load data, or when breaking down in the distributed substation of certain prognoses system, replace it to complete the function of power and load prediction.
Remote master server equalization algorithm module refers to Fig. 4
Prognoses system distributed substation core processor, for dispatching the collaborative work between each submodule, completes basic system cloud gray model and data processing.
Prognoses system distributed substation communication module is used for the communication before substation and master server, can gather local environment Monitoring Data simultaneously.
Prognoses system distributed substation memory module is for storing the historical data of nearly trimestral environment monitor, load, power.
The algorithm that prognoses system distributed substation neural network algorithm training module uses is consistent with the neural network algorithm of server, but training data is less, is generally used for the online retraining of model.
The forecast model that prognoses system distributed substation neural network algorithm training prediction module trains out according to neural network algorithm training module is weighted prediction.
Prognoses system distributed substation equalization algorithm module map 5.
Fig. 1 is system architecture schematic diagram of the present invention, and master server is used for mathematical computations and the storage of large-scale data of high complexity, and it can carry out information interaction by Ethernet switch and the distributed substation of prognoses system.The distributed substation of prognoses system can need the task of extensive computing to send to master server some, completes calculating by master server.The generated output real time data gathering blower fan, photovoltaic in micro-capacitance sensor or sub-microgrid is responsible in the distributed substation of each prognoses system, and the load data in this region.
Fig. 2 is neural network BP training algorithm flow process.Neural network algorithm needs to do two pieces thing: 1 training forecast model, 2 computational prediction results.This algorithm is the process of training forecast model.Training pattern is equivalent to the relation finding a function representation constrained input, after this model finds, as long as input data as requested, just can be predicted the outcome, Here it is " process of computational prediction result ", Fig. 2 is the process describing training pattern, model training is not predict the outcome all to need training at every turn, a model training has been got well and just can be placed on there, can with several days, some months or several years, if when finding to predict the outcome too large with actual result deviation at every turn, just need re-training model, find the relation of new input and output.Neural network adopts three layers of feed-forward framework.First be initialization weight data, each weights assignment random number, random number range is between-1 to 1.Using the input value of historical data sample as neural network, calculate the input and output of every one deck neural network respectively.After obtaining the final output of neural network, ask theoretical value and calculated value square error, if error meets pre-conditioned, then train end.If do not meet pre-conditioned, then judge maximum cycle again, if exceed maximum cycle equally also terminate training, otherwise backwards calculation every layer neuron partial gradient is revised neuron weights.Until meet the pre-conditioned of square error or maximum cycle.
Fig. 3 is the direction of prognoses system distributed substation structural representation, dotted line indication device internal control stream, the direction of solid line indication device internal data flow.The asset creation of equalization algorithm module in charge control neural network training module and Resourse Distribute, when local resource can not meet training requirement or can affect data acquisition, power load prediction, device can be trained to master server request remote opening, model file can be sent to this locality after master server completes training.
Fig. 4 is system-level load balance scheduling algorithm flow chart, whether local resource satisfies the demand, if do not met, start this algorithm, master server is after opening network connection service, circulation is intercepted substation request, once there be substation request master server to complete neural metwork training, the historical data of searching corresponding substation information and substation is in a database trained by master server.After training terminates, model file can be sent to corresponding substation.
Fig. 5 is network element internal thread-level load balance scheduling algorithm flow chart.First equalization algorithm module can check operating system version, and loads corresponding grand, initialization global variable.Algorithm can be added up how many CPU and start to monitor the occupancy of each CPU.Balance module algorithm can create a task manager and be used for task burst, task monitors, distributing system resource and recovery system resource in internal memory.If when local resource can meet all tasks, task manager can start the task after performing burst.If local resource can not meet all tasks, and this task cannot be waited for, equalization algorithm module can connect master server, is performed by this job invocation to master server.
Each subtask first can initialization task counter, ensure that this task occupying system resources time is limited, will the task of Data import, data processing be completed after this tasks carrying, and externally I/O request reading and writing of files, file lock, by ensureing the uniqueness of file read-write, prevents corrupt data.
More than show and describe ultimate principle of the present invention and principal character and advantage of the present invention.The technician of the industry should understand; the present invention is not restricted to the described embodiments; what describe in above-described embodiment and instructions just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.Application claims protection domain is defined by appending claims and equivalent thereof.
Claims (8)
1. based on a master-slave mode microgrid power load prediction system for load balancing, it comprises master server, stores for mathematical computations and data;
The distributed substation of prognoses system, for gathering the relevant real time data of local prediction, the power in prediction local zone and load data, carry out modified weight calculating simultaneously; Described master server too network switch and the distributed substation of prognoses system carries out information interaction,
Described master server comprises with lower module:
Meteorological acquisition module, for gathering the free numerical weather forecast of automatic network;
Accurate weather prognosis module is foundation according to local longitude, dimension, height above sea level, geopotential unit, simultaneously according to the geography information in micro-capacitance sensor garden and building schematic diagram, carries out three-dimensional modeling;
Database, for storing meteorological historical data and image data, and power, the load data in micro-capacitance sensor garden; Remote master server communication module is used for communicating with the distributed substation of each prognoses system, gathers the weather data on Internet simultaneously;
Communication module, based on common ethernet communication mode, uses ICP/IP protocol.
The long-range training module of neural network, for training power prediction and load forecasting model;
Equalization algorithm module, is responsible for asset creation and the Resourse Distribute of control neural network training module;
Remotely predicting module, according to general power and the total load data of the predict the outcome multiple micro-capacitance sensor of prediction or multiple sub-micro-capacitance sensor of the distributed substation of prognoses system, or when breaking down in the distributed substation of certain prognoses system, replaces it to complete power and load prediction.
2. the master-slave mode microgrid power load prediction system based on load balancing according to claim 1, it is characterized in that, the distributed substation of described prognoses system comprises:
Central processing unit, for dispatching the collaborative work between each submodule, completes basic system cloud gray model and data processing;
Prognoses system communication module, for the communication before the distributed substation of prognoses system and master server, gathers local environment Monitoring Data simultaneously;
Memory module, for storing the historical data of nearly trimestral environment monitor, load, power;
Neural network algorithm training module, for training power prediction and the online retraining of load forecasting model;
Neural network algorithm training prediction module, is weighted prediction according to the forecast model that neural network algorithm training module trains out;
Equalization algorithm module, be responsible for asset creation and the Resourse Distribute of control neural network training module, when local resource can not meet training requirement or can affect data acquisition, power load prediction, central processing unit can be trained to master server request remote opening, load forecasting model file can be sent to this locality after master server completes training;
Local prediction module, for gathering the relevant real time data of local prediction, the power in prediction local zone and load data;
Described communication module, memory module, neural network algorithm training module are all connected with central processing unit with local prediction module, described equalization algorithm module is all connected with neural network algorithm training module with neural network algorithm training prediction module, described local prediction module, neural network algorithm training module and prognoses system communication module are all connected with memory module, are convenient to the storage of data.
3. the master-slave mode microgrid power load prediction system based on load balancing according to claim 2, it is characterized in that, described local prediction module comprises photovoltaic prediction module, wind-powered electricity generation prediction module and load prediction module.
4. the master-slave mode microgrid power load prediction system based on load balancing according to claim 1, it is characterized in that, the neural network algorithm of the long-range training module of described neural network adopts the BP neural network algorithm improved, and the BP neural network algorithm of described improvement is as follows:
Right value update Δ W
ij:
Wherein, W
1jrepresent the connection weights between l neuron to an output layer jth neuron, O
ijrepresent that neural network exports, y
jrepresent desired value, v
1represent l neuron input value, θ
jrepresent threshold value, L represents neuron number.
5. the master-slave mode microgrid power load prediction system based on load balancing according to claim 1, is characterized in that, what described database adopted is Oracle or sybase database.
6., based on a master-slave mode microgrid power load forecasting method for load balancing, its method step is as follows:
(1) distributed capture prediction steps; Gather the real time data that local prediction is relevant, by the power in Neural Network Prediction local zone and load data, carry out modified weight calculating simultaneously, export budget result;
(2) meteorological acquisition step, for gathering the free numerical weather forecast of automatic network;
(3) weather prognosis step, the locality adopted according to step (1) among a small circle longitude, dimension, height above sea level, geopotential unit is foundation, carries out three-dimensional modeling; The network weather data that model obtains using step (2), as initial conditions, adopt physical equation and the meteorological change of thermodynamical equilibrium equation simulation, and after neural network algorithm correction, finally obtain accurate weather information in micro-capacitance sensor garden;
(4) general power and the total load data of micro-capacitance sensor are predicted; Predicting the outcome and general power through predicting multiple micro-capacitance sensor or multiple sub-micro-capacitance sensor and total load data according to step (1);
Described neural network algorithm comprises system-level load balance scheduling algorithm steps, training requirement can not be met when local resource or can data acquisition be affected, during power load prediction, device can to master server request remote opening load balance scheduling Algorithm for Training, master server is after opening network connection service, circulation is intercepted substation request, once there be substation request master server to complete neural metwork training, the historical data of searching corresponding substation information and substation is in a database trained by master server, model file can be sent to corresponding substation after master server completes training.
7. method according to claim 6, is characterized in that, the method for neural network algorithm is: be first initialization weight data, each weights assignment random number, and random number range is between-1 to 1; Using the input value of historical data sample as neural network, calculate the input and output of every one deck neural network respectively.After obtaining the final output of neural network, ask theoretical value and calculated value square error, if error meets pre-conditioned, then train end; If do not meet pre-conditioned, then judge maximum cycle again, if exceed maximum cycle equally also terminate training, otherwise backwards calculation every layer neuron partial gradient is revised neuron weights; Until meet the pre-conditioned of square error or maximum cycle.
8. method according to claim 6, it is characterized in that, also comprise network element internal thread-level load balance scheduling algorithm steps in described neural network algorithm, first equalization algorithm module can check operating system version, and loads corresponding grand, initialization global variable.Algorithm can be added up how many CPU and start to monitor the occupancy of each CPU; Balance module algorithm can create a task manager and be used for task burst, task monitors, distributing system resource and recovery system resource in internal memory; If when local resource can meet all tasks, task manager can start the task after performing burst; If local resource can not meet all tasks, and this task cannot be waited for, equalization algorithm module can connect master server, is performed by this job invocation to master server;
Each subtask first can initialization task counter, ensure that this task occupying system resources time is limited, will the task of Data import, data processing be completed after this tasks carrying, and externally I/O request reading and writing of files, file lock, by ensureing the uniqueness of file read-write, prevents corrupt data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410759321.XA CN104376389B (en) | 2014-12-10 | 2014-12-10 | Master-slave mode microgrid power load prediction system and method based on load balancing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410759321.XA CN104376389B (en) | 2014-12-10 | 2014-12-10 | Master-slave mode microgrid power load prediction system and method based on load balancing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104376389A true CN104376389A (en) | 2015-02-25 |
| CN104376389B CN104376389B (en) | 2017-09-15 |
Family
ID=52555284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410759321.XA Active CN104376389B (en) | 2014-12-10 | 2014-12-10 | Master-slave mode microgrid power load prediction system and method based on load balancing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104376389B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951846A (en) * | 2015-06-02 | 2015-09-30 | 国电南京自动化股份有限公司 | Microgrid short-term power and load prediction system and dynamic error classification and correction method |
| CN105159093A (en) * | 2015-10-08 | 2015-12-16 | 国电南京自动化股份有限公司 | Model-self-adapting-based energy optimal control system of microgrid and design method thereof |
| TWI559250B (en) * | 2015-10-16 | 2016-11-21 | 行政院原子能委員會核能研究所 | Real-Time Scheduling Method for Micro-grid Energy Management |
| CN106330558A (en) * | 2016-08-31 | 2017-01-11 | 哈尔滨工业大学（威海） | Controller load prediction system and method applied to software defined network |
| WO2017185387A1 (en) * | 2016-04-27 | 2017-11-02 | 北京中科寒武纪科技有限公司 | Method and device for executing forwarding operation of fully-connected layered neural network |
| CN107748934A (en) * | 2017-10-26 | 2018-03-02 | 国网新疆电力公司经济技术研究院 | Distribution electric heating load forecasting method based on improved BP algorithm |
| CN108229678A (en) * | 2017-10-24 | 2018-06-29 | 深圳市商汤科技有限公司 | Network training method, method of controlling operation thereof, device, storage medium and equipment |
| CN109242755A (en) * | 2018-08-01 | 2019-01-18 | 浙江深眸科技有限公司 | Computer vision processing server framework neural network based |
| CN109272140A (en) * | 2018-08-10 | 2019-01-25 | 国网河北省电力有限公司石家庄供电分公司 | A kind of transformer equipment runtime forecasting of cost method based on big data analysis |
| CN110232476A (en) * | 2019-05-30 | 2019-09-13 | 华中科技大学 | A kind of micro-grid load prediction technique based on deep learning |
| CN110458342A (en) * | 2019-07-26 | 2019-11-15 | 国网江苏省电力有限公司金湖县供电分公司 | One kind monitoring system and method based on improved NARX neural network microclimate |
| CN110709789A (en) * | 2017-04-10 | 2020-01-17 | Abb瑞士股份有限公司 | Method and apparatus for monitoring the condition of a subsystem within a renewable power generation device or microgrid |
| WO2020118587A1 (en) * | 2018-12-12 | 2020-06-18 | 华北电力大学扬中智能电气研究中心 | Load prediction method and device |
| CN112131178A (en) * | 2020-09-07 | 2020-12-25 | 上海电气风电集团股份有限公司 | Data processing method and device and determining method and device for fan load calculation |
| CN113837480A (en) * | 2021-09-29 | 2021-12-24 | 河北工业大学 | Impact load prediction method based on improved GRU and differential error compensation |
| CN116032190A (en) * | 2023-03-29 | 2023-04-28 | 海澜智云科技有限公司 | Motor load management system using neural network |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202126703U (en) * | 2011-06-15 | 2012-01-25 | 中国电力科学研究院 | Wind power optimization scheduling plan system |
| CN103345227A (en) * | 2013-07-02 | 2013-10-09 | 东南大学 | Micro grid monitoring and energy management device and method |
| CN103514487A (en) * | 2013-07-15 | 2014-01-15 | 国家电网公司 | Load forecasting method of power distribution network with distributed power supply |
| CN103699944A (en) * | 2013-12-31 | 2014-04-02 | 国电南京自动化股份有限公司 | Wind and photovoltaic generation power prediction system with multiple prediction modes |
| CN104008426A (en) * | 2014-05-15 | 2014-08-27 | 上海交通大学 | Distributed computing environment performance predicting method based on integrated learning |
| CN104065101A (en) * | 2014-06-26 | 2014-09-24 | 华电电力科学研究院 | Distributed microgrid energy management system |
-
2014
- 2014-12-10 CN CN201410759321.XA patent/CN104376389B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202126703U (en) * | 2011-06-15 | 2012-01-25 | 中国电力科学研究院 | Wind power optimization scheduling plan system |
| CN103345227A (en) * | 2013-07-02 | 2013-10-09 | 东南大学 | Micro grid monitoring and energy management device and method |
| CN103514487A (en) * | 2013-07-15 | 2014-01-15 | 国家电网公司 | Load forecasting method of power distribution network with distributed power supply |
| CN103699944A (en) * | 2013-12-31 | 2014-04-02 | 国电南京自动化股份有限公司 | Wind and photovoltaic generation power prediction system with multiple prediction modes |
| CN104008426A (en) * | 2014-05-15 | 2014-08-27 | 上海交通大学 | Distributed computing environment performance predicting method based on integrated learning |
| CN104065101A (en) * | 2014-06-26 | 2014-09-24 | 华电电力科学研究院 | Distributed microgrid energy management system |
Non-Patent Citations (1)
| Title |
|---|
| 吴长伟: "基于BP神经网络的负载均衡技术的研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951846A (en) * | 2015-06-02 | 2015-09-30 | 国电南京自动化股份有限公司 | Microgrid short-term power and load prediction system and dynamic error classification and correction method |
| CN104951846B (en) * | 2015-06-02 | 2018-12-28 | 国电南京自动化股份有限公司 | Micro-capacitance sensor short term power and load prediction system and error classification dynamic correcting method |
| CN105159093B (en) * | 2015-10-08 | 2018-06-29 | 国电南京自动化股份有限公司 | Microgrid energy Optimal Control System and its design method based on model adaptation |
| CN105159093A (en) * | 2015-10-08 | 2015-12-16 | 国电南京自动化股份有限公司 | Model-self-adapting-based energy optimal control system of microgrid and design method thereof |
| TWI559250B (en) * | 2015-10-16 | 2016-11-21 | 行政院原子能委員會核能研究所 | Real-Time Scheduling Method for Micro-grid Energy Management |
| WO2017185387A1 (en) * | 2016-04-27 | 2017-11-02 | 北京中科寒武纪科技有限公司 | Method and device for executing forwarding operation of fully-connected layered neural network |
| US11373084B2 (en) | 2016-04-27 | 2022-06-28 | Cambricon Technologies Corporation Limited | Apparatus and methods for forward propagation in fully connected layers of convolutional neural networks |
| CN106330558A (en) * | 2016-08-31 | 2017-01-11 | 哈尔滨工业大学（威海） | Controller load prediction system and method applied to software defined network |
| CN110709789A (en) * | 2017-04-10 | 2020-01-17 | Abb瑞士股份有限公司 | Method and apparatus for monitoring the condition of a subsystem within a renewable power generation device or microgrid |
| CN108229678A (en) * | 2017-10-24 | 2018-06-29 | 深圳市商汤科技有限公司 | Network training method, method of controlling operation thereof, device, storage medium and equipment |
| CN108229678B (en) * | 2017-10-24 | 2021-04-06 | 深圳市商汤科技有限公司 | Network training method, operation control method, device, storage medium and equipment |
| CN107748934A (en) * | 2017-10-26 | 2018-03-02 | 国网新疆电力公司经济技术研究院 | Distribution electric heating load forecasting method based on improved BP algorithm |
| CN107748934B (en) * | 2017-10-26 | 2023-06-23 | 国网新疆电力公司经济技术研究院 | Distribution network electric heating load prediction method based on improved BP neural network algorithm |
| CN109242755A (en) * | 2018-08-01 | 2019-01-18 | 浙江深眸科技有限公司 | Computer vision processing server framework neural network based |
| CN109272140B (en) * | 2018-08-10 | 2021-09-28 | 国网河北省电力有限公司石家庄供电分公司 | Big data analysis-based power transformation equipment operation period cost prediction method |
| CN109272140A (en) * | 2018-08-10 | 2019-01-25 | 国网河北省电力有限公司石家庄供电分公司 | A kind of transformer equipment runtime forecasting of cost method based on big data analysis |
| WO2020118587A1 (en) * | 2018-12-12 | 2020-06-18 | 华北电力大学扬中智能电气研究中心 | Load prediction method and device |
| CN110232476B (en) * | 2019-05-30 | 2021-12-03 | 华中科技大学 | Micro-grid load prediction method based on deep learning |
| CN110232476A (en) * | 2019-05-30 | 2019-09-13 | 华中科技大学 | A kind of micro-grid load prediction technique based on deep learning |
| CN110458342A (en) * | 2019-07-26 | 2019-11-15 | 国网江苏省电力有限公司金湖县供电分公司 | One kind monitoring system and method based on improved NARX neural network microclimate |
| CN112131178A (en) * | 2020-09-07 | 2020-12-25 | 上海电气风电集团股份有限公司 | Data processing method and device and determining method and device for fan load calculation |
| CN113837480A (en) * | 2021-09-29 | 2021-12-24 | 河北工业大学 | Impact load prediction method based on improved GRU and differential error compensation |
| CN113837480B (en) * | 2021-09-29 | 2023-11-07 | 河北工业大学 | Impact load prediction method based on improved GRU and differential error compensation |
| CN116032190A (en) * | 2023-03-29 | 2023-04-28 | 海澜智云科技有限公司 | Motor load management system using neural network |
| CN116032190B (en) * | 2023-03-29 | 2023-05-30 | 海澜智云科技有限公司 | Motor load management system using neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104376389B (en) | 2017-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104376389A (en) | Master-slave type micro-grid power load prediction system and master-slave type micro-grid power load prediction method based on load balancing | |
| Li et al. | Building's electricity consumption prediction using optimized artificial neural networks and principal component analysis | |
| Zheng et al. | Short-term wind power forecasting using a double-stage hierarchical ANFIS approach for energy management in microgrids | |
| Wan et al. | Optimal prediction intervals of wind power generation | |
| CN105046374A (en) | Power interval predication method based on nucleus limit learning machine model | |
| CN103489038A (en) | Photovoltaic ultra-short-term power prediction method based on LM-BP neural network | |
| Zheng et al. | Bio-inspired optimization of sustainable energy systems: a review | |
| CN103699944A (en) | Wind and photovoltaic generation power prediction system with multiple prediction modes | |
| CN105184416A (en) | Fluctuation wind speed prediction method based on particle swarm optimization back propagation neural network | |
| US10474177B2 (en) | System and method for stability monitoring, analysis and control of electric power systems | |
| Sengar et al. | Ensemble approach for short term load forecasting in wind energy system using hybrid algorithm | |
| CN107092989A (en) | The Forecasting Methodology and equipment of short-term wind-electricity power | |
| CN104331737A (en) | Office building load prediction method based on particle swarm neural network | |
| Jiang et al. | Day‐ahead renewable scenario forecasts based on generative adversarial networks | |
| CN104657584A (en) | Lorenz-system-based wind speed prediction method | |
| CN112101626A (en) | Distributed photovoltaic power generation power prediction method and system | |
| Alharbi et al. | Short-term wind speed and temperature forecasting model based on gated recurrent unit neural networks | |
| CN112801357B (en) | Solar radiation quantity prediction method, device, equipment and storage medium | |
| CN105844350A (en) | Short period wind power prediction system based on covariance preferable combination model | |
| Pliuhin et al. | Smart Grid technologies as a concept of innovative energy development: initial proposals for the development of Ukraine | |
| CN112821456B (en) | Distributed source-storage-load matching method and device based on transfer learning | |
| Zeiler et al. | Integral BEMS controlled LVPP in the smart grid: An approach to a complex distributed system of the built environment for sustainability | |
| CN114552570A (en) | Offshore wind power prediction management system | |
| CN107391689A (en) | Wind power output abnormal data reconstructing method based on big data technology | |
| CN112001518A (en) | Prediction and energy management method and system based on cloud computing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |